Author: Vincent Diepeveen
Date: 20:34:55 07/06/00
Go up one level in this thread
On July 06, 2000 at 22:24:18, Dann Corbit wrote: >On July 06, 2000 at 22:03:26, Robert Hyatt wrote: > >>On July 06, 2000 at 19:41:31, Vincent Diepeveen wrote: >> >>>Hello, >>> >>>this is a post i just did at crafty list after a long number of mails about >>>a project called CAP which people tend to believe they can use it for >>>creating their openingsbook. Some warnings are on their place, when you >>>want to let such a book compete with commercial books. >>> >>>I first showed that CAP goes completely wrong already quite soon in opening. >>>Like 1.d4,d5 2.c4,e5 3.dxe5 there best move is d4 all other moves you can >>>go home and rest but first resign the game. >> >>This is a bogus title for a thread. CAP isn't right in all cases. Neither >>is "most popular move". Nor "move suggested by Kasparov (or Anand or Karpov >>or any other GM)". >> >>CAP data _is_ good and reliable. You simply have to use it as one of several >>decision-making ideas. Not _the_ decision making idea. Weighted with other >>things like frequency, learning, even simple positional evaluation, CAP data >>can contribute to a better move choice. Used by itself, it can also contribute >>to a bad choice. As can even a top GM's suggestion(s). >> >>If you don't want to use the data, don't. But I don't see any reason to >>suggest that others should not try it. They might have a better idea than >>you do (or than I do). >> >>Dismissing something just because you don't see (yet) how to use it is not >>a good way to make progress. Had I dismissed bitmaps so quickly, I would have >>missed something that has turned out to be very interesting and useful. > >This is the ideal time to use CAP as a punching bag. We have this big pile of >raw data and it has not been studied very carefully. It should be fairly simple >to poke a few holes into a straw man like "Abandon your opening books and use >CAP data since it is so great." > >I don't think anyone has advocated that, at this stage of the game. > >I think that possibly the most interesting results will come from an analysis of >the 60 million plus positions analyzed at fast time control. I have some >machines with large memory here, and I might just load the whole shooting match >as a giant tree and minimax the whole thing. > >Who knows what interesting thing might fall out of it? > >At any rate, what we have is raw information, and as such, I think it's pretty >good information. [In my completely unbiased and aloof scientific estimation] >;-) > >I will say that in my opinion, CAP is in the "Kitty-Hawk" stage right now. It >does not look like it would be of much value for transportation, but you never >know. > >And (in my continued estimation) if you have _no_ opening book (perhaps half of >the winboard programs are in that sad state of affairs) then CAP data is >certainly better than *nothing*. And, combined with a few thousand SuperGM >games, you can make a decent opening book. It won't have the quality of a >commercial tool, but someday even that might be generated automatically. > >I know Vincent's opinion is that we won't make it past the primorial ooze stage, >but I think I am patient enough to wait and see. I see CAP as a classical scientific way to waste time on something which is insignificant compared to the goal you might want to achieve. If your goal is only to produce positions with an evaluation of a chessporgram added then you clearly succeed in that goal. If your goal is to do something that's useful for those who want to create openingsbooks, then i hope i have warned them for the fact that CAP data is saying nothing in itself. A score of +0.8 or -0.8, what does it tell us? It only tells us that a certain program's summation after minimaxing is this score, it is just too inaccurate to blindfoldedly go on the score of the program. When comparing it to a state of the art hand tailored book where you have produced a lot of games computer versus computer, then the CAP data is completely insignificant compared to these computer-computer data, assuming that data gets also hand interpreted. In the end only interpretation by an expert can lead to the awaited result. The CAP data+statistical information about GM games in itself cannot be used as a decision criteria if you want to get to a level where you want to beat hand created books by experts, wanting to be that it sure eats time making something like that. It sure is better to have something instead of nothing, but we all already have something and that something are a lot of books and analyzes. Vincent
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.